Using the Yunnan evolutionary population synthesis (EPS) models with and without binary interactions, we present the luminosity of Hα recombination line (LHα), the luminosity of [O ii] λ3727 forbidden line doublet (), the ultraviolet (UV) fluxes at 1500 and 2800 Å (Li, UV) and far-infrared flux (LFIR) for burst, S0, Sa–Sd and Irr galaxies, and present the calibrations of star formation rate (SFR) in terms of these diagnostics.
By comparison, we find that binary interactions lower the SFR versus LHα and SFR versus conversion factors by ∼0.2 dex. The main reason is that binary interactions raise the UV flux (shortwards of the Lyman limit) of the stellar population (SP) in the age range 6.7 < log t/yr < 8.4 and thus more ionizing photons are present in the nebula. Moreover, binary interactions do not significantly vary the calibrations of SFR in terms of Li, UV. This is because binary interactions raise the flux at 1500 Å of the SP in the range 8.75 < log t/yr < 9.2 and the maximal difference is about 1 dex. In addition, binary interactions have little effect on the flux at 2800 Å. At last, the calibration of SFR from LFIR is almost unaffected by binary interactions. This is caused by the fact that binary interactions almost do not affect the bolometric magnitudes of SPs.
We also discuss the effects of initial mass function (IMF), gas-recycle assumption and EPS models [including GISSEL98 (Galaxy Isochrone Synthesis Spectral Evolution Library), BC03, starburst99, POPSTAR and pégase models] on these SFR calibrations. Comparing the results obtained by using the Salpeter (hereafter S55) IMF with those by using the Miller & Scalo (hereafter MS79) IMF, we find that the SFR versus LHα and SFR versus conversion factors obtained by using the S55 IMF are greater by 0.4 and 0.2 dex than those by using the MS79 IMF for the Yunnan models with and without binary interactions, respectively. The SFR versus Li, UV and SFR versus LFIR conversion factors by using the S55 IMF are larger by an amount of 0.2 dex than the corresponding ones by using the MS79 IMF. The inclusion of gas-recycle assumption only lowers these SFR calibrations at faint SFR. Moreover, comparing the results when using different EPS models, we find that the differences in the SFR versus LHα and SFR versus conversion factors reach ∼0.7 and 0.9 dex, the difference in the SFR versus LFIR conversion factor reaches 0.4 and 0.8 dex, and the differences in the SFR versus Li, UV conversion factors reach 0.3 and 0.2 dex when using the S55 and non-S55 IMFs (including Chabrier, Kroupa–Aarseth–Hurley, Kroupa–Tout–Gilmore and Miller–Scalo IMFs, partly caused by the difference in the IMF), respectively. At last, we give the conversion coefficients between SFR and these diagnostics for all models.